Financial Crime Investigation Copilot
Outcome
Compress alert-to-decision time for financial crime investigators by assembling transaction chains, KYC evidence, and watchlist matches into a narrative that an investigator confirms before any suspicious activity report is filed.
| Portfolio attribute | Value |
|---|---|
| Primary persona | AML investigator |
| Primary aggregate | InvestigationCase |
| Cosmos DB containers | 2 |
| Grounding data | transactions, KYC, alerts, cases, watchlists |
| Development estimate | USD 60-220 monthly dev estimate |
| Runtime options | Python FastAPI and C# ASP.NET Core |
Productivity hypothesis
These measures reflect the workflow, evidence, decision boundary, and industry risk for this use case.
| Impact type | Measure | Conservative internal assumption | Customer measurement |
|---|---|---|---|
| Productivity | Case assembly cycle time | 8-15% reduction | (baseline median case-assembly minutes - assisted median minutes) / baseline median minutes * 100 |
| Quality or risk | Investigation rework | 5-10% reduction | (baseline reopened or corrected investigation rate - assisted rate) / baseline rate * 100 |
Important: Conservative internal pilot assumptions only; not public benchmarks, commitments, or expected customer results. Validate against a comparable baseline or matched control cohort with mandatory human review.
Rationale: Transactions, KYC, alerts, cases, and watchlists support faster case assembly, not an unvalidated claim of improved crime detection.
Pilot guardrail: Preserve analyst judgment, evidence lineage, access controls, and formal case-handling procedures.
Customer validation plan
- Baseline the current workflow with representative users and tasks.
- Run the accelerator on the same task mix with human review retained.
- Measure case assembly cycle time and investigation rework with the documented formulas.
- Replace internal assumptions with customer-observed results before a business case.
Why Azure Cosmos DB
This AI solution needs an operational data plane, not only a vector index:
- Hybrid vector and full-text retrieval combines semantic recall with exact identifiers, codes, names, and domain language.
- Hierarchical partition keys and metadata predicates keep tenant and domain boundaries in the normal query path.
- Operational JSON, source chunks, workflow and agent state, feedback, and audit records remain together for low-latency reads and governed writes.
- Change feed reacts to source and operational updates without polling.
- Autoscale throughput, configurable consistency, and multi-region distribution support variable demand and enterprise availability.
Domain workflow

| Stage | Actor | Input → output | Approval |
|---|---|---|---|
| 1. Triage and consolidate related alerts | Triage agent | Alert stream with subject identifiers → Consolidated case with linked alerts | Policy controlled |
| 2. Resolve parties and beneficial ownership | Entity resolution service | Subject identifiers and KYC records → PartyProfile set with ownership graph | Policy controlled |
| 3. Reconstruct flow of funds | Transaction analytics agent | Transaction ledger for the review window → TransactionChain with hop detail and totals | Policy controlled |
| 4. Draft the investigation narrative | Narrative agent | Chains, screening hits, and typology hypothesis → Cited NarrativeDraft with recommended disposition | Policy controlled |
| 5. Decide disposition and authorize filing | Investigator and compliance officer | Narrative draft and evidence pack → Closed case or authorized regulatory filing | Human required |
Technical architecture

Cosmos DB data model
| Container | Hierarchical partition key | Documents | Search paths |
|---|---|---|---|
investigationCases | /tenantId, /caseId | InvestigationCase, AlertRecord, TransactionChain, ScreeningHit, NarrativeDraft | /adverseMediaVectorEmbedding, /narrativeVectorEmbedding, /adverseMediaVector, /narrativeVector |
partyMasterProfiles | /tenantId, /partyId | PartyProfile | /profileNarrativeVectorEmbedding, /profileNarrativeVector |
Query contract
| Query | Container | Scope | Purpose |
|---|---|---|---|
| Case evidence assembly | investigationCases | single-partition | Read the full evidence set for one case in a single partition-aligned query. |
| Regulatory clock watchlist | investigationCases | cross-partition-bounded | Bounded cross-partition control-plane query: surface cases approaching the statutory reporting deadline. It cannot carry the second partition key because deadline risk is only visible when every open case partition is compared, so it runs as a tenant-scoped scheduled sweep off the interactive path, bounded by open cases whose regulatory due date falls inside a five-day horizon, executed with MaxItemCount paging and continuation tokens. |
| Unresolved screening hits | investigationCases | cross-partition-bounded | Bounded cross-partition control-plane query: list high-score sanctions or adverse media matches still awaiting disposition across the case book. It cannot carry the second partition key because the sanctions queue is owned by a central team rather than a single case, so it runs as a tenant-scoped scheduled sweep off the interactive path, bounded by undecided dispositions above the analyst score threshold, executed with MaxItemCount paging and continuation tokens. |
| High-value structuring scan | investigationCases | cross-partition-bounded | Bounded cross-partition control-plane query: find chains whose aggregate value crosses the structuring review threshold. It cannot carry the second partition key because structuring is detected by comparing chains that belong to different cases, so it runs as a tenant-scoped scheduled sweep off the interactive path, bounded by a 30-day transaction window and a monetary review threshold, executed with MaxItemCount paging and continuation tokens. |
Python and C# services route each document type to its declared container and hierarchical partition key. ETag-based writes protect mutable aggregates, while append-only records and bounded cross-partition queries preserve scale.
Azure deployment

Clone the source repository, authenticate with Azure, and run:
Set-Location accelerators/banking-capital-markets/financial-crime-investigation-copilot
az login
./Deploy.ps1 -Runtime python -Location eastus2
Change python to csharp to deploy the .NET implementation.
Explore the implementation
- Launch signed-in demo
- Implementation article
- Domain data model
- Workflow and API contract
- Python application
- C# application
- Domain container Bicep
- Platform Bicep
- Cost and configuration
- Source repository